Include burden curation from relevant publications

[ireneisdoomed]

We want to keep adding burden results from relevant publications.

Background

The GWAS Catalog puts together a list of publications with gene-level results that are hence susceptible of being added to our gene burden evidence. The original spreadsheet is updated every so often and is here, and it currently has ~150 publications. Since we don't have capacity to curate all of them, we want to find a way to prioritise the most impactful ones. I have created another slightly modified version that adds the publication journal and title to help assessing what could be worth bringing into the Platform https://docs.google.com/spreadsheets/d/1EQrDPGMB2l4lkOWvzBxu-Ek7jgC7edYK5XPuGFdLbcE/edit?usp=sharing This will be probably useful in the future, the code is here https://docs.google.com/spreadsheets/u/1/d/1EQrDPGMB2l4lkOWvzBxu-Ek7jgC7edYK5XPuGFdLbcE/edit?usp=sharing

Tasks

Together with @buniello we have identified publications that are interesting to cover.

[ireneisdoomed]

WES data from SCHEMA consortium associates 10 genes with schizophrenia

Reference: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9805802/#SD2

The paper describes a study that used ultra-rare coding variants to identify 10 genes that are significantly associated with schizophrenia.

Authors aggregated case-control and trio-based de novo mutations for gene discovery.The analysis included 24248 cases, 97322 controls, and 3402 parent-probad trios.

They looked at PTVs and missense variants, categorizing them as Class 1 (MPC > 3), Class 2 (MPC 2 - 3), and de novo variants. They also looked at synonymous variants to build the expected null distribution. On top of that, they performed meta-analyses to analyze the data collectively, which is how most of the associations were found.

As cohorts they used exomes from the SCHEMA and gnomAD consortia. 24,248 individuals with schizophrenia and 97,322 controls from seven continental populations.

Results are mainly reported in Table 1, which I have expanded we ST5. We can extract 14 pieces of evidence from this study:

• 10 as the result of the meta analysis (so we won't have directionality)
• Association with TRIO, XPO7, HERC1 and SETD1A showed statistical significance with class 1 variants. Here we do have effect sizes.

[ireneisdoomed]

Depression burden results

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10897433/

• 7 definitions of depression
• 17 sets of variants: PTV, MPC > 2, 2 ≥ MPC > 1, 1 ≥ MPC > 0, other missense and synonymous variants in pLI ≥ 0.9 genes (6 groups); PTV, 2 ≥ MPC > 1, 1 ≥ MPC > 0, other missense and synonymous variants in pLI < 0.9 genes (5 groups); PTV, MPC > 2, 2 ≥ MPC > 1, 1 ≥ MPC > 0, other missense and synonymous variants for all genes (6 groups).
• They don't report any relevant genes. All gene-level results are uploaded to Zenodo. There's a file per each type of depression x collapsing set. Complicated (https://zenodo.org/records/10511823)

I wouldn't spend time on this.

[ireneisdoomed]

Prostate cancer in African samples

Reference: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10697980/#MOESM1033/

African ancestry is a significant risk factor for prostate cancer and advanced disease. The study used rare coding variants from WES to identify genes significantly associated with prostate cancer.

• P value threshold in P < 5E−04
• Cohort of 798 people of African ancestry, 451 cases
• Statistical Analyses: SKAT-O analysis (no OR) of protein altering (nonsynonymous, slicing and nonsense) variants
• Genes showing association (MTG1, H3C1, MBP) have no (or very little) info about their role in prostate cancer